Starting Soon: Soil Sampling and Decision Making Using

1 Starting Soon: Soil Sampling and Decision Making Using Incremental Sampling Methodology (ISM) - Part 2

Incremental Sampling Methodology Technology Regulatory and Guidance Document (ISM-1, February 2012) - https://itrcweb.org/GuidanceDocuments/ISM-1_2012_with_Clarifications.pdf

Download PowerPoint file• Clu-in training page at https://clu-in.org/conf/itrc/ISM/

• Under “Download Training Materials”

Using Adobe Connect• Related Links (on right)

Select name of link

Click “Browse To”

• Full Screen button near top of page

Follow ITRC

No associated notes.

2

2

Incremental Sampling Methodology Technology Regulatory and Guidance Document (ISM-1, February 2012)

Welcome – Thanks for joining this ITRC Training Class

Sponsored by: Interstate Technology and Regulatory Council (www.itrcweb.org) Hosted by: US EPA Clean Up Information Network (www.cluin.org)

Soil Sampling and Decision Making Using Incremental Sampling Methodology (ISM)

Part 2 – Implement, Assess, and Apply

Web-Based Document at: http://www.itrcweb.org/ISM-1/

Follow ITRC

When sampling soil at potentially contaminated sites, the goal is collecting representative samples which will lead to quality decisions. Unfortunately traditional soil sampling methods don't always provide the accurate, reproducible, and defensible data needed. Incremental Sampling Methodology (ISM) can help with this soil sampling challenge. ISM is a structured composite sampling and processing protocol that reduces data variability and provides a reasonable estimate of a chemical's mean concentration for the volume of soil being sampled. The three key components of ISM are systematic planning, field sample collection, and laboratory processing and analysis. The adequacy of ISM sample support (sample mass) reduces sampling and laboratory errors, and the ISM strategy improves the reliability and defensibility of sampling data by reducing data variability.ISM provides representative samples of specific soil volumes defined as Decision Units. An ISM replicate sample is established by collecting numerous increments of soil (typically 30 to 100 increments) that are combined, processed, and subsampled according to specific protocols. ISM is increasingly being used for sampling soils at hazardous waste sites and on suspected contaminated lands. Proponents have found that the coverage afforded by collecting many increments, together with disciplined processing and subsampling of the combined increments, yields consistent and reproducible results that in most instances have been preferable to the results obtained by more traditional (e.g. discrete) sampling approaches.This 2-part training course along with ITRC's web-based Incremental Sampling Methodology Technical and Regulatory Guidance Document(ISM-1, 2012) is intended to assist regulators and practitioners with the understanding the fundamental concepts of soil/contaminant heterogeneity, representative sampling, sampling/laboratory error and how ISM addresses these concepts. Through this training course you should learn:-- basic principles to improve soil sampling results-- systematic planning steps important to ISM-- how to determine ISM Decision Units (DU)-- the answers to common questions about ISM sampling design and data analysis-- methods to collect and analyze ISM soil samples-- the impact of laboratory processing on soil samples-- how to evaluate ISM data and make decisions

In addition this ISM training and guidance provides insight on when and how to apply ISM at a contaminated site, and will aid in developing or reviewing project documents incorporating ISM (e.g., work plans, sampling plans, reports). You will also be provided with links to additional resources related to ISM.The intended users of this guidance and training course are state and federal regulators, project managers, and consultant personnel responsible for and/or directly involved in developing, identifying or applying soil and sediment sampling approaches and establishing sampling objectives and methods. In addition, data end users and decision makers will gain insight to the use and impacts of ISM for soil sampling for potentially contaminated sites.Recommended Reading: We encourage participants to review the ITRC ISM document (http://www.itrcweb.org/ISM-1/) prior to participating in the training classes. If your time is limited in reviewing the document in advance, we suggest you prioritize your time by reading the Executive Summary, Chapter 4 "Statistical Sampling Designs for ISM," and Chapter 7 "Making Decisions Using ISM Data" to maximize your learning experience during the upcoming training classes.

ITRC (Interstate Technology and Regulatory Council) www.itrcweb.org

Training Co-Sponsored by: US EPA Technology Innovation and Field Services Division (TIFSD) (www.clu-in.org)

ITRC Training Program: [email protected]; Phone: 402-201-2419

3

3

Housekeeping

Course time is 2¼ hours

This event is being recorded

Trainers control slides• Want to control your

own slides? You can download presentation file on Clu-in training page

Questions and feedback• Throughout training:

type in the “Q & A” box

• At Q&A breaks: unmute your phone with #6 to ask out loud

• At end of class: Feedback form available from last slide Need confirmation of your

participation today? Fill out the feedback form and check box for confirmation email and certificate

Copyright 2017 Interstate Technology & Regulatory Council, 50 F Street, NW, Suite 350, Washington, DC 20001

Although I’m sure that some of you are familiar with these rules from previous CLU-IN events, let’s run through them quickly for our new participants.

We have started the seminar with all phone lines muted to prevent background noise. Please keep your phone lines muted during the seminar to minimize disruption and background noise. During the question and answer break, press #6 to unmute your lines to ask a question (note: *6 to mute again). Also, please do NOT put this call on hold as this may bring unwanted background music over the lines and interrupt the seminar.

Use the “Q&A” box to ask questions, make comments, or report technical problems any time. For questions and comments provided out loud, please hold until the designated Q&A breaks.

Everyone – please complete the feedback form before you leave the training website. Link to feedback form is available on last slide.

4

4ITRC (www.itrcweb.org) – Shaping the Future of Regulatory Acceptance

Host organization Network

• State regulators All 50 states, PR, DC

• Federal partners

• ITRC Industry Affiliates Program

• Academia• Community stakeholders

Follow ITRC

Disclaimer

• Full version in “Notes” section

• Partially funded by the U.S. government

ITRC nor US government warranty material

ITRC nor US government endorse specific products

• ITRC materials copyrighted –see usage policy

Available from www.itrcweb.org

• Technical and regulatory guidance documents

• Internet-based and classroom training schedule

• More…

DOE DOD EPA

The Interstate Technology and Regulatory Council (ITRC) is a state-led coalition of regulators, industry experts, citizen stakeholders, academia and federal partners that work to achieve regulatory acceptance of environmental technologies and innovative approaches. ITRC consists of all 50 states (and Puerto Rico and the District of Columbia) that work to break down barriers and reduce compliance costs, making it easier to use new technologies and helping states maximize resources. ITRC brings together a diverse mix of environmental experts and stakeholders from both the public and private sectors to broaden and deepen technical knowledge and advance the regulatory acceptance of environmental technologies. Together, we’re building the environmental community’s ability to expedite quality decision making while protecting human health and the environment. With our network of organizations and individuals throughout the environmental community, ITRC is a unique catalyst for dialogue between regulators and the regulated community.

For a state to be a member of ITRC their environmental agency must designate a State Point of Contact. To find out who your State POC is check out the “contacts” section at www.itrcweb.org. Also, click on “membership” to learn how you can become a member of an ITRC Technical Team.

Disclaimer: This material was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof and no official endorsement should be inferred.

The information provided in documents, training curricula, and other print or electronic materials created by the Interstate Technology and Regulatory Council (“ITRC” and such materials are referred to as “ITRC Materials”) is intended as a general reference to help regulators and others develop a consistent approach to their evaluation, regulatory approval, and deployment of environmental technologies. The information in ITRC Materials was formulated to be reliable and accurate. However, the information is provided "as is" and use of this information is at the users’ own risk.

ITRC Materials do not necessarily address all applicable health and safety risks and precautions with respect to particular materials, conditions, or procedures in specific applications of any technology. Consequently, ITRC recommends consulting applicable standards, laws, regulations, suppliers of materials, and material safety data sheets for information concerning safety and health risks and precautions and compliance with then-applicable laws and regulations. ITRC, ERIS and ECOS shall not be liable in the event of any conflict between information in ITRC Materials and such laws, regulations, and/or other ordinances. The content in ITRC Materials may be revised or withdrawn at any time without prior notice.

ITRC, ERIS, and ECOS make no representations or warranties, express or implied, with respect to information in ITRC Materials and specifically disclaim all warranties to the fullest extent permitted by law (including, but not limited to, merchantability or fitness for a particular purpose). ITRC, ERIS, and ECOS will not accept liability for damages of any kind that result from acting upon or using this information.

ITRC, ERIS, and ECOS do not endorse or recommend the use of specific technology or technology provider through ITRC Materials. Reference to technologies, products, or services offered by other parties does not constitute a guarantee by ITRC, ERIS, and ECOS of the quality or value of those technologies, products, or services. Information in ITRC Materials is for general reference only; it should not be construed as definitive guidance for any specific site and is not a substitute for consultation with qualified professional advisors.

5

5

Meet the ITRC Trainers

Mark BruceTestAmerica Inc.North Canton, OH330-966-7267mark.bruce@

testamericainc.com

Tim Frederick USEPA Region 4Atlanta, [email protected]

Jay ClausenU.S. Army Cold Regions

Research and Engineering Laboratory

Hanover, NH603-646-4597Jay.L.Clausen@

usace.army.mil

Michael StrohMissouri Department of

Natural ResourcesJefferson City, MO573-522-9902michael.stroh@

dnr.mo.gov

Michael Stroh is a senior project manager with the Missouri Department of Natural Resources in Jefferson City, Missouri. He joined the department in 1997 to oversee site assessment and cleanup activities in the state’s Voluntary Cleanup Program before joining the state' Superfund Site Assessment group in 2002. Before joining MDNR, Michael worked as an analytical chemist and section supervisor forenvironmental laboratories in the private sector. He has also worked as a chemist for the US Geological Survey at a mobile laboratory on the Chesapeake Bay, and as a biohydrologist for The Nature Conservancy in Arlington, Virginia. Michael joined the ITRC Incremental Sampling Methodology (ISM) team at its inception in 2009 and has served as a chapter lead in drafting the technical and regulatory guidance document. Michael earned a bachelor's degree in biology from Truman State University in Kirksville, Missouri in 1985 and a master's degree in environmental science from the Indiana University School of Public and Environmental Affairs in Bloomington, Indiana in 1994.

Jay Clausen is a Physical Research Scientist with the US Army Corps of Engineers, Engineer Research and Development Center, Cold Regions Research and Engineering Laboratory (CRREL) in Hanover, New Hampshire. Since 2005, Jay has conducted research on the application of the Incremental Sampling Methodology (ISM) to military ranges containing energetics and metals in particulate form as well as the fate-and-transport of lead, tungsten, and energetics. Jay routinely presents at conferences and publishes peer-reviewed journal articles, and government reports. Prior to joining CRREL, Jay worked for 14 years in the environmental consulting field for AMEC as a Senior Hydrogeologist/Project Manager and Lockheed Martin focused on Department of Defense and Department of Energy issues. Jay earned a bachelor’s degree in geology from the University of Nebraska in Omaha in 1987 and a master’s degree in the geosciences from the University of Maine in Orono in 1990. Jay is currently working on PhD in the Natural Resources and Earth System Science program at the University of New Hampshire focused on the application of ISM techniques to soils containing metals. Jay is a certified professional geologist in the states of Kentucky, New Hampshire, Texas, and Washington and a certified professional hydrogeologist in the state of Washington.

Mark Bruce is a Corporate Technical Director for TestAmerica. He has experience in environmental monitoring of metals, wet chemistry, volatile and semivolatile organics including sample preparation and analysis since 1979. He has participated in the development of several EPA methods. His responsibilities include method development, evaluation, implementation and troubleshooting. He provides technical support to TestAmerica clients and internal staff. He has worked with incremental sample processing techniques since 2003. He has been active on the ITRC Incremental Sampling Methodology team since it started in 2009 and particularly involved with the development of the laboratory guidance. Mark earned a Ph.D. in Analytical Chemistry from the University of Cincinnati in 1984.Tim Frederick is a human health risk assessor with EPA Region 4 in Atlanta, GA. He provides risk assessment, toxicology, and other technical support to EPA’s project managers on a wide variety of sites in the Southeast. Tim has worked in the Region’s Superfund Division since 2005 and serves on several national workgroups. Prior to joining EPA, Tim provided risk assessment and toxicology technical support to EPA as a contractor from 1996 to 2005, and he worked as an environmental consultant in Miami, Florida from 1990 to 1996. Tim has presented nationally at technical conferences, and he has provided guest lectures at local universities. Tim has been a member of the ITRC’s Incremental Sampling Methodology team since it began in 2009. Tim earned a bachelor's degree in psychology from the University of Miami in Miami, Florida in 1989 and a master’s of public health in environmental and occupational health from Emory University’s Rollins School of Public Health in Atlanta, Georgia in 1996.

6Soil Sampling Data –What Do We All Need?

Accuracy

Reproducibility

Defensibility

….but how do we get it?

Incremental Sampling Methodology (ISM)

…..may be your answer…..

Picture Reference: http://www.swrcb.ca.gov/rwqcb2/brownfields.shtml

Set stage – context with Day 1

6

7

7

ISM Goal – A Representative Sample!

The primary objective of ISM is to obtain

a representative sample having constituents

in exactly

the same

proportions

as the entire

decision unit.

Representative subsampling

Speaker Bullets

•Cannot analyze the whole decision unit, i.e. put the whole unit in a jar

•Sampling Error should be limited/managed to obtain sample representative of the field/DU

•a representative subsample of the fraction of interest determined during systematic planning is collected in the laboratory and will be presented in the next module

•Sampling Error is managed with ISM through the collection of 30 or more increments of correct shape and adequate mass

•Field implementation and sampling activities should be performed by ISM trained personnel

Supplemental Information

Section 5.1 Introduction

Picture Reference: http://www.swrcb.ca.gov/rwqcb2/brownfields.shtml

8Incremental Sampling Methodology (ISM) Process

WHAT IS ISM?

This slide shows quite succinctly all the steps involved in developing triplicate analytical results for a particular sampling unit. Triplicate results are what we are recommending for a number of reasons – largely related to addressing the heterogeneity we talked about earlier, and being able to make some statistical estimate of variability. Triplicates will allow a calculation of the 95% UCL that many of you – particularly risk assessors - are interested in.

Remember –

•minimum of 30 increments (20 – 60g/increment)

•3 replicates in order to get a 95 UCL

•Sample size should be ~1 kg (600 – 2,500 grams).

8

9

ISM Document and Training Roadmap

Plan

Principles

Systematic Planning

Statistical Design

Implement

Field Implementation

Lab Processing

Assess Making Decisions

Introduction Challenges/Opportunities

Application

Part 1

Part 2

Section 2

Section 3

Section 4

Section 5

Section 7

Section 8

Section 6

ISM DocumentTraining Module

ISMOpportunities


9

10 ISM Part 1 – SummaryPrinciples, Systematic Planning, and Statistical Design

Reduce Sampling Errors• Heterogeneity Rules!

Plan, Plan, Plan• Involve the entire team

• Know your site

• Know your objectives

• Focus your decisions

Design for Confidence• The mean is the goal!

• Collect replicates to calculate UCL

Plan

Principles

Systematic Planning

Statistical Design

The last training included Modules on:

Soil and Principles – Be aware of issues related to heterogeneity and sampling errors

Systematic Planning - Involve the entire team, regulators, consultants, responsible parties in critical elements (e.g. conceptual site model, establish sampling objectives and decision units.) Sampling objectives should drive your sampling design, and the scale of decision making should align with sampling objectives.

Statistical Design - Provides the statistical foundation and describes why ISM provides a reasonable mean, describes a good ISM sampling design, and informs you how ISM provides 95% UCL.

10

11ISM Part 2 PreviewImplement, Assess, and Apply

Implement


Lab Processing

AssessMaking

Decisions

Application

Collect an

ISM Sample

Match Lab Process to Analytes and Objectives

Decision Mechanisms and Data Evaluation

Where to Apply ISMISM

Opportunities

?


11

12Field Implementation Learning Objectives

Learn how to:

Collect an ISM sample • Understand the similarities and differences

between surface and subsurface ISM sampling

ITRC, ISM-1, Section 5

• Consider issues specific to non-volatile and volatile ISM sampling

• Implement and collect ISM replicate samples

Speaker Bullets

With correct up front planning and coordination, ISM field implementation can be as simple as putting dirt in a bucket (or bag).


ITRC, ISM-1, Section 5

12

13

Key Presentation Topics

Sampling design

Sampling tools

ISM surface/subsurface sampling• Cores and subsampling

Specific contaminant of

concern (COC)

considerations• Non-volatile and volatile

ISM replicates

Speaker Bullets

•Brief review of sampling designs that were presented as part of Day 1, Section 4

13

14

ISM Field Sampling Implementation

ITRC, ISM-1, Section 5.1, Figure 5-1

Speaker Bullets

Figure 5-1 from the technical regulatory document shown as an example only. The audience is not expected to read this flowchart (small font); provided only as an example to illustrate the multiple decisions that will need to be made and implemented.

Multiple step process that all should be discussed/planned for during systematic planning

Technical regulatory document will assist with these steps


ITRC, ISM-1, Section 5.1, Figure 5-1

14

15

15

Sample Collection Components

Decision Unit (DU) sampling design • Simple random sampling

• Random sampling within a grid

• Systematic random sampling

Sampling tools• Core shaped

• Adequate diameter

Mass• Increment mass

• Sample mass

ITRC, ISM-1, Section 4.3.4.2

Speaker Bullets

•Sampling Design - 3 basic sampling designs

•The appropriate sample support for tools to address soil sampling theory to limit error

•Adequate mass collected to limit error


Section 4.3.4.2 Effects of sampling pattern

16

16

Sampling Designs

Simple RandomRandom within Grids

Systematic Random

Increments

ITRC, ISM-1, Section 4.3.4.2 & Section 5.3.1, Appendix A1

Speaker Bullets

•As a brief review, just as with discrete sampling, a variety of sampling methods may be implemented with ISM sampling.

• The advantages/disadvantages were presented as part of the Section 4 of the Day 1 training. Additional information is found in Chapter 4 of the ISM document.

•Simple Random Sampling, Systematic Random Sampling, and Systematic Grid Sampling all yield unbiased estimates of the mean. All are appropriate for ISM sampling, choice would be project specific

•“unbiased” meaning every soil particle has an equal chance of being collected/analyzed

•Note that “random” in this context does not mean wherever the sampling team feels like taking a sample, and that a formal approach to determining the random increment locations must be used

•These are 2 dimensional examples, however, recall that DU’s have a depth/3 dimensional component

•Recommended minimum 30 increments per DU regardless of sampling design

•The systematic random sampling is generally more efficient to implement in the field; only one (initial) random location is determined (3 initial random locations for replicates); and

• It ensures relatively even spatial distribution of increments across the DU, therefore avoids the appearance of over- or under-representation of sub-areas within a DU, (as may occur with SRS - Simple Random Sampling); reduce/control the size of any potential unsampled area (“hotspot”) to be detected/not detected.


ITRC, ISM-1, Section 4.3.4.2 & Section 5.3.1, Appendix A1

17

17Florida Case Study: Decision Unit (DU) Identification

Identify DU in the field• Use typical environmental site investigation procedures

• Examples Survey

GPS

Swing ties

ITRC, ISM-1, Section 9.3 & Appendix C, Section C.3

Decision Unit (~1/4 acre)

Speaker Bullets

•DU(s) determined during systematic planning

•Standard environmental site investigation procedures employed to delineate DU in the field

•The photograph is from the Florida case study that was introduced as you recall during the ISM Day 1 training. Former golf course with arsenic as the contaminant of concern. The DU is a ¼ acre residential lot – assumed exposure area of potential future residential development.

•Chapter 9, Section 9.3 and Appendix C, Section C.3 for Former Golf Course Field Demonstration of ISM


ITRC, ISM-1, Section 9.3 & Appendix C, Section C.3for Former Golf Course Field Demonstration of ISM

18

18

Increment Locations

Identify increment locations in field• Utilize similar site investigation tools

ITRC, ISM-1, Section 5.3.1

Speaker Bullets

•Sampling design, number of increments, increment locations are determined during systematic planning (Day 1 training)

•Typically 30 or more increments per DU

•Similar tools are utilized to identify increment locations; e.g. survey grid, GPS, sampling flags

•Example of transformer pad and residential backyard DUs

•Increment locations are flagged in these examples. May assist inexperienced ISM practitioners, more experienced ISM samplers it may not be necessary.



19

19Florida Case Study: Increment Field Determination

Speaker Bullets

•It may or may not be necessary to determine/document the “exact” location of each increments, depending on criteria and DQOs specified during systematic planning. Check with regulatory authority.

•Generally not necessary, however regulations or policy, public perception, education, etc. may require.

•For the Florida Case study, specific increment locations were not marked/flagged. The locations (relative to the gridpoints, e.g., 2 ft east and 3.5 ft south) were recorded in the SAP (Sampling and Analysis Plan) that was prepared prior to the study. A tape measure was used to measure the first few locations exactly, the distance was converted to “heel-to-toe steps”, and then this method was used for all of the remaining increments.

20

20

Sampling Tool Considerations

Criteria - shape• Cylindrical or core shaped increments

• Minimum diameter required – based on particle size (soil fraction) of interest

e.g., core diameter >16 mm

ITRC, ISM-1, Section 5.2

Speaker Bullets

•Core increments should be collected

•Minimum diameter of core determined per fraction of interest; e.g. generally soil is considered <2mm fraction, minimum core diameter of 16 mm. See section 5.2 of ISM document for additional information on calculation/recommended core diameter.

•Larger fraction of interest (or particle size) requires a larger diameter core sampler

•Complete core increments should be collected across the depth of the DU (cohesive/noncohesive)

•Commercially available tools (next slides)



21

21

Additional Considerations

Decontamination• Not necessary within DU (including replicates)

Sampling tool• Appropriate for matrix and contaminant of interest


Speaker Bullets

•Tools can be used without decontamination between increments within DU (including replicates); all increments are part of the same sample, cross contamination is not an issue.

•Collected directly into appropriate sample container (Ziploc lower right) not hands, only for example purposes.

•Example – Multi-Incremental Sampling Tool (MISTTM)

•Other commercially available tools (next slides)



22

22

Sampling Tool Examples

Soft Surface Soil

Source: Courtesy http://www.jmcsoil.com/index.htmlhttp://fieldenvironmental.com/evc-incremental-sampler.php

Speaker Bullets

•Examples of commercially available tools.

•Other tools, e.g. agricultural soil samplers, may also be applicable to ISM.

•Top: EVC Incremental Sampler (kit); Bottom: JMC Backsaver Handle and Soil Tubes

23

Alternate Sampling Tools

Hard-Packed or Gravelly Soil

Speaker Bullets

•Alternate tools may be applicable depending on matrix

23

24

24Florida Case Study:Field Sampling

Speaker Bullets

•Florida Golf Course Example – Every job needs at least one supervisor.

•Surface core increments collected in plastic 5 gallon bucket

25

25Florida Case Study:“Low Tech” Sampling Tools

Speaker Bullets

Low tech alternatives, if site conditions amenable

Arsenic only COC; PVC would not interfere/contaminate analysis of COC (phthalates)

Left pounding PVC section into ground; right PVC increments from DU, all increments then combined into 1 ISM sample

26

26

Adequate Sample Mass

Criteria – mass (non-volatile)• Recommended mass per increment: 20-60 grams

• Final ISM samples: generally 600-2,500 grams

Ms = ρ • n • Ds • π • (q / 2)2

Ms – targeted mass of sample (g)

Ds – increment length (cm)

n – number of increments

ρ - soil or sediment density (g/cm3)

q - diameter of sample core (cm)


Speaker Bullets

•All parameters should be determined during systematic planning

•Generally, a minimum of 30 increments should be collected for each DU



Example of bulk ISM sample ~2 kg

27

Individual core samples combined to prepare an ISM sample for each DU

Subsurface Decision Units (DU)

-1.5’

-0.5’

-3.0’

-5.0’

-10’

DU-1

DU-2

DU-3

DU-4

30 Borings (minimum recommended)Core Increments

not to scale

Speaker Bullets

•Same DU determination (3 dimensional volume), sample design, number of increments, etc. as surface sampling

•Subsurface DUs presented in Module 3

•DUs designated at depth to provide vertical resolution of contamination within DU and optimize remediation.

27

28

28

Preferred increment – entire core interval Core subsampling alternatives

1. Core wedge2. Core slice

Subsurface Sampling Considerations


Speaker Bullets

•Entire core increment preferred; logistically subsampling the cores may be necessary.

•Cores may be subsampled in field and combined to prepare ISM sample for designated DU layer



29

29

Core Wedge

Continuous wedge removed from entire length of targeted DU interval for 100% coverage


e.g., wedge width >16 mm

Speaker Bullets

•“Wedge” as pictured is taken across entire core interval/DU depth

•The simplest approach is to split the core in half, vertically along the axis, reducing the increment mass by half. Alternatively, a single wedge of soil is taken from the entire length of the targeted depth interval.

•Wedge width requirement per fraction of interest (sample support) the same as core diameter (e.g. 16 mm for 2mm soil); larger fraction of interest requires wider wedge.



30

30

Core Slice

Core Slice removed from randomly selected interval length of targeted DU depth


Speaker Bullets

•Randomly selected core “slice” collected from core interval as increment

•Least preferred subsampling method. Increased bias over previous options due to incomplete depth interval collection/representation



31

31

Field Processing for Non-Volatiles

ISM sample processing in a controlled laboratory environment is recommended to reduce error

Field processing may be applicable if project specific DQOs can be met


Speaker Bullets

•Each processing step introduces some amount of sampling error; sample processing in a controlled laboratory environment recommended to reduce this error

•Generally follows same process as laboratory, i.e. sieving, spread evenly in thin layer, collect multiple systematic random increments to form subsample ( i.e., 2 dimensional (2D) Japanese slabcake subsampling)

•Subsampling not recommended for particulate contaminants (energetics, metals), entire sieved fraction should be ground and subsampled

•Field subsampling triplicate samples recommended to evaluate error



32

32Florida Case Study: Non-Volatile ISM Sample Logistics

Final ISM samples: typically 600-2,500 grams or more

• Containers, storage, shipping

Laboratory

• Facilities and equipment for correct processing and subsampling

Speaker Bullets

•All logistical considerations should be determined during site-specific systematic planning

•Large samples, alternate containers, temperature and preservation considerations, etc.

•Samples were transferred from 5-gallon buckets to Ziploc bags, however, larger sample support may require storage and shipment of larger (5 gallon) samples to the laboratory.

•Laboratory coordination and communication essential

33

33

ISM Volatile Sampling Tools

Core type sampler

Typical for VOC soil sampling per SW846 5035A

ITRC, ISM-1, Section 5.4.2 Source: Courtesy www.ennovativetech.com

Speaker Bullets

•Core increments on smaller (typical 5 gram) scale

•Standard VOC collection for high level, field preserved soil samples; SW846 Method 5035A, Section 8.2.2. Difference is multiple core increments placed into single adjusted volume of methanol



Left) Core N’ One; Center) Terra Core Sampler; Right) Easy Draw Syringe & PowerStop Handle

34

34

ISM Volatile Samples – Subsurface

Each core represents an increment from the DU Layer

Subsamples of increment collected from each core to prepare bulk sample (e.g., 5g plug every 5cm, total subsample mass 20-60g to produce bulk ISM sample)

Speaker Bullets

•Core, wedge, slice increments impractical for VOCs

•Representative VOC sampling of subsurface core

•Potential VOC loss; increments should be collected and preserved a quickly as possible from opened core

•Photo on right is an example of a core with VOC plugs removed at ~2 inch intervals

35

35

Methanol

Soil

VOC preservation and analysis• Increments are extruded from sampler directly into

volume of appropriate container with predetermined methanol

• Methanol preserved sample submitted to laboratory

• Note shipping restrictions/ requirements

ISM Volatile Sample Logistics

ITRC, ISM-1, Section 5.4.2, Figure 5-11

Speaker Bullets

•Close laboratory coordination required for the appropriate sample bottles and methanol volume

•Shipping logistics/restrictions may apply; Guidelines for the transportation of a solvent such as Methanol can be found in Department of Transportation (DOT), 49 Code of Federal Regulations (CFR), Part 172, List of Hazardous Substances (ground) and Reportable Quantities. Shipments via air transport may also be required to adhere to International Air Transport Association Dangerous Goods Regulations (IATA DGR) (air).


ITRC, ISM-1, Section 5.4.2, Figure 5-11

See section 5.4.2 for additional information and options for methanol

36

36

Replicates Recommended

Increments collected from alternate random locations

• Independent samples, not “splits”

Minimum 3 replicate set for statistical evaluations

Additional replicates may be necessary depending on contaminant heterogeneity and project specific DQOs


Speaker Bullets

• Recommended to evaluate spatial heterogeneity and ensure reliable estimate of the mean

• Required for any statistical evaluation, e.g. 95% UCL of the mean

•Refer to Day 1 training presentation of ISM replicates, estimation of mean and coverage

•Refer to Chapter 4 of the ISM document for the statistical basis/evaluation of ISM replicates



37

Replicate Spacing and Collection

Decision Unit

Replicate Increment Spacing

Decision Unit

Sample Collection

R1 R2 R3Replicate 1

Replicate 2

Replicate 3ITRC, ISM-1, Section 5.3.5

Speaker Bullets

•Field replicate example for systematic random sampling design; different random starting location, “serpentine pattern” if collected in succession

•The collection order (of increments/replicates) does not matter; what is most efficient in the field

•Submitted blind to laboratory



37

38

38

Field Replicates – Simple Example

Replicate 3

Replicate 2

Replicate 1

Collecting the Samples

Speaker Bullets

•ISM replicate locations; stockpile sampling with grid demarcated with string

39

Replicate/Sampling Reminders

Replicates • What type

• How many

• Where/when will they be collected

• How will they be evaluated

“Homogenizing” or mixing not necessary• Laboratory processing and subsampling (following

module) designed to attain representative analytical sample

Speaker Bullets

•What type – DU and/or field subsampling replicates

•How many – 3 or more depending on DQOs

•Where – replicate increment location

•When – simultaneously within grid or consecutively

39

40

Field Implementation Summary

Determined during Systematic Planning• Sampling design

• Adequate sampling tools

• ISM surface/subsurface sampling logistics Subsurface cores and subsampling

• Specific contaminant of concern (COC) considerations

• Non-volatile and volatile

• ISM replicates

Speaker Bullets

•All items should be determined during systematic planning and implemented in the field to obtain representative samples to submit to the laboratory for additional processing and/or analysis.

•That concludes this module.

40

41


Plan

Principles

Systematic Planning

Statistical Design

Implement


Lab Processing



Application

Part 1

Part 2

Section 2

Section 3

Section 4

Section 5

Section 7

Section 8

Section 6


ISMOpportunities

Speaker bullets

•Focus – Match lab process

•To analytes

•And end data use

Narrative

Thank you Earl (or Jay). We have now learned how important planning is. Followed by diligent work to collect representative samples in the field. Now we need to maintain that level of effort at the laboratory. Next we are going to discuss how to match the right lab processes to the analytes of interest and the end uses of the data.

41

42Laboratory ProcessingLearning Objectives

Learn how to:

Match process options to analytes and data objectives

Manage sample moisture

Select/reduce particle size

Collect subsamples for analysis

Apply Quality Assurance

Examine options for lab certification

Speaker bullets

•Focus – manage sample moisture, particle size selection and reduction and subsampling

•Ripple effects – digestion and extraction subsamples

•QC adaptations

Narrative

We will focus on matching the processing options to the analytes and data needs. Laboratory support for ISM primarily focuses on the preliminary soil moisture management, particle size selection and particle size reduction and analytical subsampling. There are some ripple effects for sample digestion and extraction. Quality assurance processes should also be adapted to cover these new sample handling steps. Last we will examine the options for lab certification.

42

43

Real Life ISM has Choices

?

ITRC, ISM-1, Section 6, Figure 6-1

Speaker bullets

•Many process choices

Narrative

Real life ISM processing has a variety of options. Sample processing decisions should be made upfront during project planning. We will be discussing those options and how to choose amongst them.

There is no universal process that is optimal for all analytes and matrices. Although the general dry, grind, subsample and analyze is the most common it is not appropriate in all cases. There are other options.

43

44

44

Analyte-Matrix Driven Options

Pick the right option• More representative subsamples

• Better precision

Pick the wrong option• Poor and unknown bias

Speaker bullets

•Pick the option based on analyte / matrix

•Pick right – representative subsample / results / better decisions

•Pick wrong – poor & unknown bias / risk wrong decision

Narrative

The specific combination of analytes and matrix will determine which processing options are the best to meet the project specific data quality objectives. Pick the right option and the result is more representative subsamples, leading to more representative results with better precision to facilitate better decisions. Pick the wrong option and you end up with poor and unknown bias in the final results, potentially leading to the wrong decision.

45Include Lab Processing in Project Planning

Lab


Speaker bullets

•Project planning, include lab processing

Narrative

Include the lab processing options in the initial project planning discussions. Please don’t assume that just feeding the lab dirt samples will automatically produce good data.

Supplemental information

See Section 6.1.1

45

46

Define the Analytes

Volatile organics

Energetics

Metals, Hg

PCBs

Organochlorine pesticides

Phenoxy acid herbicides

Petroleum hydrocarbons

Semivolatile organics

Other

Speaker bullets

•Determine analyte groups

•List individual analytes

•Energetics such as explosives and propelents

•Specific analytes >> optimal process

Narrative

All analyte groups as well as the individual analytes within each group must be clearly defined and listed. The optimal sample handling options depend on the specific analytes. These differences will be highlighted in the discussion that follows.

46

47

Coordinate VOC Sampling & Analysis

Use methanol preservation• Methanol transport

• Bottle sizes (large, medium, small)

Analytical sensitivity limitations• Higher reporting limits

• Selected Ion Monitoring GC-MS Short analyte lists


Speaker bullets

•30 mL transport

•Bottle sizes

•Higher reporting limits

•SIM GC-MS for short lists

Narrative

Careful coordination between lab and field crew is usually needed because the lab supplies custom methanol preserved bottles designed for the number of increments in each specific decision unit. Recall that Department of Transportation regulations restrict shipment of containers with more than 30 mL of methanol.

Use of the methanol extraction solvent facilitates combining many increments into a single analysis, but does limit the theoretical analytical sensitivity when compared to in-vial purge methods using aqueous solutions. Some data users consider the methanol extraction to be more efficient at recovering VOCs even though the calculated reporting limits tend to be higher.

Selected ion monitoring GC-MS analysis can produce lower reporting limits when working with short analyte lists.


Section 6.2.1

47

48Florida Case Study:Contaminant of Concern

Arsenic • From liquid applied pesticides

As

Periodic Table of Elements

Speaker bullets

•Arsenic

•Liquid pesticide

Narrative

Arsenic was the contaminant of concern for the Florida golf course study. It had been applied in liquid form as a pesticide.

48

49

Symbol Key

Good effect

Bad effect

Result or statistic gets larger in value

Result or statistic gets smaller in value

Speaker bullets

•Coded arrows

•Solid green - good

•Striped red – bad

•Upward pointing - larger value

•Downward pointing - smaller value

Narrative

Through the remainder of the lab processing section you will see several coded arrows. Solid green arrows mean the effect on the analytical results is good. Striped red arrows means the effect is bad. An upward pointing arrow means that the numerical statistic or result is larger. A downward pointing arrow means that the numerical statistic or result is smaller. The size of the arrow represents the magnitude of the change.

49

50

Lab Processing Roadmap

Lab Processing

Sample Conditioning

Particle Size Reduction

Splitting and Subsampling


50

51

Air drying• Room temperature – most common

• Ventilation hood

• Goal: Crushable agglomerates

• Consider volatilization losses Boiling point

Binding to soil particles

Potential for Loss Table– Naphthalene

– Acenaphthene

– Benzo[a]pyrene

Use other options when drying not appropriate

Condition the Sample


Speaker bullets

•Air dry

•Ventilation

•Low boiling – volatilization losses

•Binding to particles

•Loss risk table

•Other conditioning options available when air drying not appropriate

Narrative

The most common sample conditioning step is air drying, usually at room temperature in a ventilation hood. The goal is to produce dry crushable dirt clods that work well with the mechanical processes that can follow. Air drying can produce low recovery of analytes with lower boiling points that are not strongly bound to the soil particles. The guidance has a loss risk table that summarizes the relative risk of loosing various semivolatile organics and energetics during air drying. For example, the risk of loosing naphthalene is large, acenaphthene is moderate and benzo[a]pyrene is small. Diesel and mercury have a wide variety of components or species with a range of volatilities. If lower boiling analytes are of primary interest, it might be necessary to avoid air drying.

When working with surface soils that have been exposed to air for many months or years after the potential contaminant release, the likelihood of volatilizing additional contaminants is small.


See Section 6.2.2.3

51

52Florida Case Study:Air Drying Samples

Arsenic • High boiling arsenic species

• Volatilization loss not expected

ITRC, ISM-1, Section 9, Appendix C

Speaker bullets

•Case study example, look for this

•Highlight the process decisions for lab process module

•Data shown in next module

•Air drying

Narrative

Since only high boiling arsenic species were expected in these surface soil samples for the golf course study, air drying was selected as the initial sample processing step.


Chapter 9 Appendix C

52

53

Define Terms: Grinding

Generic term for soil disaggregation or milling

The grinding type or equipment must be specified to select a particular laboratory process

Speaker bullets

* Grinding defined

Narrative

Grinding is a generic term that includes soil disaggregation or milling. The specific grinding technique or equipment must be specified to select a particular lab process. By itself the word grinding is not specific enough for a lab to know what process choices are needed to support the project DQOs.

53

54

Define Terms: Disaggregating

Breaking all the soil clumps into individual small particles, but keeping the small pebbles and hard crystalline particles intact


Speaker bullets

•Disaggregation defined

•Compared to milling

•Higher std dev of subsampling

•Lower measured metals concentration

Narrative

Disaggregating means breaking the soil clumps into individual small particles but keeping the small pebbles and other hard crystalline materials intact. This can be accomplished using a bladed coffee chopper or gently crushing the dirt clods on a sieve. Even a hand mortar and pestle can be used gently to break up the clods without milling the stones.


See the end of Section 6.2.2.3

54

55

Picture from USACE-Alan Hewitt

Define Terms: Milling

Complete particle size reduction of all soil components including hard crystalline materials to a defined maximum particle size (e.g. < 75 µm)


Speaker bullets

•Milling defined

•Compared to disaggregation

•Lower std dev

•Higher measured metals concentration

Narrative

Milling on the other hand is complete particle size reduction of all soil components including the hard crystalline materials to a defined maximum particle size.


See Section 6.2.2.5

55

56Florida Case Study:Particle Size Reduction

Disaggregation and sieving• Nugget effect expected to be small

Contaminant exposure sprayed as a liquid

Mill• Puck mill

Comparison study planned

Speaker bullets

•Disaggregate and #10 sieve

•Sample split for milled and unmilled processing

Narrative

For the golf course study, after air drying the samples were disaggregated and sieved (#10, < 2mm). The nugget effect was expected to be small. Each sample was split with half for puck milling and the remainder kept unmilled for a comparison study.

56

57


Lab Processing

Sample Conditioning




57

58To Mill or Not to Mill? (Particle Size Reduction)

Recommended

• Crystalline particles, fibrous threads, paint chips

• Energetics, metals

Strengths

• Reduces variability

• Reduces subsampling error

• Facilitates mixing

• Improves precision

Picture from USACE-Alan HewittITRC, ISM-1, Section 6.2.2.5

Speaker bullets

•Recommended

•When nugget effect is significant

•Energetics

•Metals

Narrative

“To mill or not to mill” depends on the analytes, nugget effect and data quality objectives. Milling is recommended for energetics and metals when present in the sample as crystalline particles or fibrous threads. The particle size reduction associated with milling reduces variability. It reduces the subsampling error and improves precision.


See Section 6.2.2.5

58

59

Not recommended

• Volatile, thermally labile, increased “availability”

• Examples Monochloro PCBs,

reactive SVOCs, decane, elemental mercury

• Limitations Analyte losses

Metals contamination

Potential high bias to metals risk assessment (pebbles)

To Mill or Not to Mill

If uncertain, do milled & unmilled


Speaker bullets

•Not recommended

•Organics

•Thermal stability

•Low boiling

•Increase metal availability

•Metals contamination

Narrative

Milling is generally not recommended for organics such as PCBs, Pesticides, hydrocarbons, semivolatile organics and some metals because some individual analytes in these groups can be lost through volatilization or thermal decomposition. Also, these analytes are less likely to be present as crystalline particles. The use of a metal mill can produce selected metal contamination of the sample. For some risk assessment purposes releasing the metals from inside the pebbles to produce higher measured concentrations is a potential high bias for the results.


See Section 6.2.2.5

59

60

How Best to Mill

Puck mill or ring and puck mill• “Stable” energetics

Ball mill

Mortar and pestle

Consider• Analytes

• Concentration of interest

• Mill materials

• Particle size needed

Picture from USACE-Alan Hewitt

Example mills, other types are possible as well


Speaker bullets

•Puck mill

•Ball mill

•Mortar and pestle

•Compare analyte to mill materials

Narrative

If one chooses to mill a soil sample there are several options. The puck mill has been demonstrated to be effective for energetics (Method 8330B) and some metals. The puck mill produces the most consistently small particles among the options, thus reducing the standard deviation of replicate results the most. The ball mill is a less expensive and less rigorous option that is still suitable for many applications. It produces a moderate improvement in precision. The mortar and pestle, either automated or manual, typically produces some precision improvement. When choosing the milling equipment consider the analytes, concentration of interest, mill materials and particle size needed. This is not an exclusive list of mills, other equipment can be used, if it meets data quality needs.


See Section 6.2.2.5

60

61Florida Case Study:Results Confirm Milling Not Needed

Disaggregation and sieving• Nugget effect expected to be small

Contaminant exposure sprayed as a liquid

Mill• Puck mill

Results confirm milling not neededfor this part of site• Small precision improvement with milling

• No change in mean concentration

Speaker bullets

•Only a small difference between milled and unmilled results.

•Milling not needed

•Due to liquid applied COC

Narrative

For the golf course, the results from the comparison of unmilled and milled sample results showed that there was essentially no change in the mean concentration results and only a small improvement precision. These results confirm that milling was not needed for this site where the arsenic contaminant of concern was distributed as a liquid and the nugget effect was small.

61

62


Lab Processing

Sample Conditioning




62

63

Dry Splitting Options

Rotary sectorial splitter


Speaker bullets

•Rotary sectorial splitter

•Best precision

•More expensive

•Less available

Narrative

The automated rotary sectorial splitter produces the most consistent and representative sample splits, thus providing the best precision. In this technique the sample flows from the hopper on top of the machine and is equally distributed into the small sample containers on the rotating sample collector.

Sectorial splitter subsampling can substantially minimize bias and improve precision.


See Section 6.2.2.7

63

64

Subsampling Options

2-Dimensional Japanese Slabcake

Dry

Wet


Speaker bullets

•2 D slabcake

•Lower cost than sectorial splitter

•Pretty good representativeness

•Wet or dry sample

•Systematic Random design

•All increments combined = analytical subsample

Narrative

The 2 dimensional Japanese slabcake frequently provides acceptable subsample representativeness at a lower cost than the sectorial splitter. This process is a miniaturized version of what takes place in the systematic random field sample collection process. The wet or dry processed sample is spread evenly in rectangular slabcake and divided into grids as determined in project planning. The default is 30. The analyst removes a small increment from a random location in the first grid. Subsequent increments are collected from the same location in the other grids. All increments are combined to form the subsample for digestion or extraction so the size of the increment must be appropriate for the number of increments and the target subsample size.

2D slabcake subsampling can minimize bias and improve precision.


See Section 6.2.2.7

64

65

Subsampling Tools

Square straight-sided scoops for dry non-cohesive soil

Speaker bullets

•Non-cohesive soil

•Blunt scoop with sides

•Not round scoop

Narrative

Proper increment collection from the 2 dimensional slabcake should use a blunt end scoop with straight sides so that the particles from the top, middle and bottom of the slabcake are all sampled equally. Rounded spoons, scoops and spatulas should not be used because they discriminate against the particles on the bottom of the slabcake and can bias the subsample. The proper tool improves subsample representativeness and reduces bias.

65

66Florida Case Study:Choose Subsampling Process

2-D Slabcake Subsampling• Lower cost than sectorial splitter

• More representative than “dig a spot”

Narrative

The 2 dimensional slabcake was chosen as the best balance between cost and representativeness for the golf course study.

66

67

Why Use Large Subsamples?

Larger particles• Produce larger errors or require larger subsamples

0

50

100

150

200

0 1 2 3 4 5

Particle size (mm)

%R

SD

1 g

5 g10 g

30 g


Speaker bullets

•Large particles >> more variable results

•Reduce variability – larger subsamples

•See Day 1 training about the heterogeneity of soil for the causes

Narrative

When particle size reduction through milling is not appropriate or available, the use of larger subsample sizes can reduce the variability and produce better precision for replicate samples. For example, with metals analysis of 1 g samples, %RSD > 100% is expected when large particles are present. Precision can be improved significantly without milling by increasing the subsample size to 5 or 10 g. During the Part 1 training the heterogeneity of soil touched on this correlation between variability and particle size.


See Section 6.3.3 and ASTM Standard D6323 Sec A1.1

67

68Florida Case Study:Nugget Effect Minimal

2 g subsamples on disaggregated aliquots

2 g subsamples on milled aliquots

Low heterogeneity expected • Confirmed through replicates

Narrative

The golf course study used 2 g subsamples for both the disaggregated and milled soil samples since the nugget effect was expected to be small. The results from this case study will be discussed in the next module.

68

69

Laboratory Quality Control Measures

Laboratory equipment blanks• Limited clean matrices

Laboratory control samples (LCS) and matrix spikes• Practicality of large scale spiking in kg samples

High cost

Limited availability

• Introduced post ISM processing into subsample

Subsampling replicates

Speaker bullets

•QC

•Blanks – no single clean matrix for all analytes

•LCS – matrix + cost

•MS - cost

•Replicates - triplicates

Narrative

Quality control measures include equipment or method blanks, laboratory control samples, matrix spike samples and subsampling replicates. There is no one single clean matrix for use with all analytes that mimics the characteristics of soil. Large scale LCSs and matrix spikes currently have significant cost associated with them and may not always be practical. Currently it is more common to introduce these QC measures after the preliminary ISM processing is complete. Subsampling replicates of field samples are suggested as a good way to assess laboratory reproducibility. Triplicates are the most common choice.


See Section 6.4

69

70Florida Case Study:Challenges with “Blank” Samples

Ottawa sand method blank attempted for milling• Metals content of the sand was too variable

Standard preparation batch QC• No laboratory control sample or matrix spike

through ISM processes

Narrative

The Florida golf course study attempted to use Ottawa sand as a milling equipment blank material, but in the end it was determined that the arsenic content in the as-received material was too variable. Standard preparation batch QC was used. No laboratory control samples or matrix spikes were run through the various ISM processes.

70

71

Verify Laboratory Certification

National Environmental Laboratory Accreditation Program (NELAP)

Non-NELAP state accreditation

Agency-specific accreditation • DoD Environmental Laboratory Approval Program


Speaker bullets

•Certification options from these

•NELAP

•State

•Agency (e.g. DoD)

Narrative

Lab certification is available through the usual avenues; the National Environmental Laboratory Accreditation Program run by the NELAC Institute, non-NELAP states and agency specific accreditation such as the DoD Environmental Laboratory Approval Program.

.


See Section 6.4.1

71

72

Cite Reference Methods

Collecting and Processing of Representative Samples For Energetic Residues in Solid Matrices from Military Training Ranges• USEPA SW-846 Method 8330B, Appendix A

http://www.epa.gov/osw/hazard/testmethods/pdfs/8330b.pdf

Metals in Solid Matrices • USACE research effort

• Planned SW-846 Method 3050 - Update V?


Speaker bullets

•Reference methods

•8330B

•3050

•Certification usually based on analysis method, secondarily extraction of digestion and subsampling ignored, but now covered for ISM

Narrative

At this point there are few references methods to cite. EPA Method 8330B for explosives is the best known and most widely used and recognized. The US Army Corps of Engineers is working on a modification of EPA metals digestion Method 3050. Hopefully this will make it into Update V for SW-846.


See Section 6.4.1

72

73

Use Alternate References

ASTM D6323 Standard Guide for Laboratory Subsampling of Media Related to Waste Management Activities • ASTM 2003

Guidance for Obtaining Representative Laboratory Analytical Subsamples from Particulate Laboratory Samples• Gerlach 2003

Laboratory Standard Operating Procedure


Speaker bullets

•General references

•ASTM

•EPA

•Lab SOP

Narrative

There are two general references related to subsampling and particle size reduction, one from ASTM and the other from the EPA. In the absence of a specific reference method some states will certify based on a specific laboratory standard operating procedure.


See Section 6.4.1

73

74

Lab Process “Big Rocks”

Match processes

and analyte needs

Manage sample

moisture

Subsample with

correct tools and process

Disaggregate To mill or not mill?

Speaker bullets

•Big Rocks

•Match analyte and processes

•Manage moisture

•Disaggregate

•Mill or not

•Subsample

Narrative

In closing, there are five foundational “big rocks” that must be addressed for lab processing of ISM samples. First, match the processes and the analyte needs. Second, manage sample moisture according to these analyte needs. Third, disaggregate samples by breaking up the soil clods. Fourth, answer “To mill or not to mill?” and choose the appropriate process. Fifth and last use the right subsampling process with the right tool to collect the most representative subsample to run through the sample digestion or extraction process.

The next module will be presented by Tim Frederick from the EPA Region 4 (Jay Clausen from the US Army Corps of Engineers)

74

75

75

Q&A Break

Decision Unit (~1/4 acre)

Follow ITRC


76


Plan

Principles

Systematic Planning

Statistical Design

Implement


Lab Processing



Application

Part 1

Part 2

Section 2

Section 3

Section 4

Section 5

Section 7

Section 8

Section 6


ISMOpportunities


76

77Making Decisions:Learning Objectives

Learn how to:

Use ISM data to make decisions

Evaluate data

• Identifying sources of error

• Quantify error

• Interpret error

• Isolate sources of error


77

78

Decision Mechanisms

Making Decisions Using ISM Data

Making Decisions

Data Evaluation

Speaker bullets

•Introduce decision mechanisms (not necessarily different from the way we make decisions with discrete data, but some wrinkles with ISM data that must be considered).

•Decision mechanisms chosen BEFORE data collection

•Data Evaluation – tools we use to assess whether we have adequately addressed project goals

•Data evaluation occurs AFTER data collection

78

79

Making Decisions

Decision Mechanism (DM)• Structured approach to making decisions

• Identified and agreed upon during Data Quality Objective (DQO) process

• 6 common types of DM

Bullet 1: ITRC Definition: an algorithm or protocol that results in the decision about a potential contaminant of concern or for a decision for a volume of media.

Each decision mechanism has strengths, weaknesses, and assumptions. In some cases, agency requirements will dictate the approach to be used – not unique to ISM, but there are some wrinkles to ISM that should be considered

Bullet 2: understand what the decision is and how it will be determined; avoid confusion –make a priori decisions; It is critical during strategic planning that parties agree up front;

Bullet 3: indicate that these are examples; you should come up with site-specific criteria (not all DMs will be acceptable in all states, scope before hand)

79

80DM 1: Compare One ISM Result to Action Level

Decision Unit Action Level

Single Result

ITRC, ISM-1, Section 4.2.1 and Section 7.2.1

Speaker Bullets - The simplest decision mechanism is to compare a single ISM result to an Action or Screening level. There are some limitations to this approach, however. -- verifying something that you expect (well above or below) – Refer to Module 4 (4.2.1) and Module 7 (7.2.1)

Narrative -

This approach might be acceptable when the predicted mean concentration is much greater than, or much less than, the action level. In this situation, the ISM sample provides confirmation of what is already strongly suspected—that the DU clearly passes or fails. However, when the ISM sample result is close to the action level, this decision mechanism is unreliable and decision errors in both directions are possible (i.e., concluding that the DU fails when the average concentration is in fact below the action level, or vice versa). This approach may not be acceptable to regulatory agencies for decision making. Acceptability of this approach will depend largely upon the objectives of the sampling.

80

81DM 2: Compare Average ISM Result to Action Level

Decision Unit Action Level

Mean of Replicates


Narrative - In this decision mechanism, a number of replicate ISM samples are collected in the field from the same DU. These replicates provide a measure of the variability of the entire sampling, preparation, and analytical process. The mean concentration of the replicates is calculated and compared to the action or screening level identified during systematic planning.

The mean concentration from replicate samples is likely to be closer to the true mean of the DU than the result from a single ISM sample (see Ch 4), and could, therefore, be considered to provide a more reliable estimate of the mean. There is no assurance, however, that the actual mean concentration has not been underestimated.

Consequently, this decision mechanism would not be useful in circumstances where project objectives dictate that the mean concentration estimate must be conservative (e.g., EPC values in most human health risk assessments). *****Some regulators may ask that error be calculated****accuracy vs precision*****

81

82Florida Case Study: Decision Mechanism (DM) 2

Discreten = 30

Incr-30 n = 3

Incr-100n = 3

DU 2 4.2 5 5.2

DU 3 7.5 10.5 9.5

Mean arsenic concentrations (mg/kg)

Narrative – This table presents the data from two DUs from the Florida pilot project (see Appendix C for more information). Please note, however, that the 30 discrete samples collected in the quarter-acre DU is more than would be generally collected on a typical site.

In this investigation, incremental sampling methods resulted in higher contaminant concentrations – this is a noteworthy example for those who worry that incremental sampling systematically dilutes concentrations

82

83

Decision Unit

Action level or risk assessment

95%UCL

DM 3: Calculate 95%UCL then Compare to Action Level or Use for Risk Assessment


• Typically for comparison of a 95% UCL to an action level, use in a risk assessment, or wherever a conservative estimate of the mean is required (where 95% UCL would be used as an exposure point concentration)

• Need at least three replicates – results improve (closer to the mean) w/ more replicates

Narrative - Project objectives may specify that the estimate of the mean concentration provided by ISM sampling must be health protective, meaning that there is low chance of underestimating the actual mean concentration within the DU, such as for use in a baseline risk assessment.

Calculation of a 95% UCL for ISM data requires a minimum of three ISM samples. This is fewer than is required for discrete data sets to yield reliable 95% UCL values. Additional replicates increase the performance of the mean estimate (i.e., provides a 95% UCL closer to the actual mean), and although this increases the cost, it may be worthwhile if the site is relatively heterogeneous and the result is anticipated to be close to the action level.

Section 4.2.2 link to the calculator

83

84Florida Case Study: Decision Mechanism 3: (DU 1)

Discreten = 10

(mg/kg)

Incr-30n = 3

(mg/kg)

Incr-100n = 3

(mg/kg)

Mean 2 1.8 1.7

Std Dev 1.4 0.08 0.03

95UCL 3.0 2.0 1.8

Florida Action Level: 2.1 mg/kg

Arsenic Data (mg/kg)

Speaker Bullets - Means are similar; variability (std dev) much lower for incremental samples; variability affects 95% UCL; in this case – method used determines the decision to be made

84

85

Decision Unit

Comparison

Background

Mean

& S

.D. M

ean

& S

.D.

DM 4: Compare to Background

ITRC, ISM-1, Section 4.4.3.3 and Section 7.2.4

Narrative:

ISM results are not suitable for point-by-point comparison with UTLs generated from discrete sample background data because ISM and discrete datasets have fundamentally different characteristics. If background and site data are both generated using ISM, comparisons of central tendencies (e.g., medians) can be made using hypothesis testing, but statistical power to detect differences will be low due to the limited number of replicates in most ISM data sets. Therefore caution is urged when comparing site to background. The best approach is to use graphical analysis to compare the ISM site data and ISM background data.

See Section 7.2.4 for more information

85

86

DM 5: Combining Decision Units

Action Level

DU average andWeighted average


Combining results from two or more small DUs to estimate the overall mean concentration in a larger combined DU is advantageous when the data must support more than one decision (e.g., overlapping exposure units for ecological and human health receptors). In this example, The decision Units A, B, C, and D (red squares) represent 1/4 –acre residential exposure units. A decision must be made on each of these units. However, an ecological receptor may range over the entire site, and a separate decision may need to be made on the larger “super DU” (blue square) Both of these project goals could be accomplished by pooling the data of the smaller DUs to calculate a weighted average for the “super DU.”

The formulas for calculating weighted averages are presented in Chapter 4 of the ITRC Tech Reg. These formulas take into account the size of the smaller DUs, weighting their contribution to the larger DU accordingly. The ability to combine DUs extends vertically as well as horizontally; that is results from DUs from different soil depths can be combined if needed to meet sampling objectives

86

87DM 6: Extrapolation to Unsampled Areas

Action Level

Sampled Decision Unit

Unsampled Decision Unit

Extrapolate

ITRC, ISM-1, Section 4.4.4.2 and Section 7.2.6

Speaker Bullets – Not unique to ISM ---fundamental assumption made with this mechanism is that the distributions of contaminant concentrations in the unsampled areas are essentially the same as in the sampled areas, in the absence of data to verify the assumption, the confidence in the assumption is subjective.

This DM (decision mechanism) involves extrapolating data from a sampled DU to an unsampled DU. The results from the sampled DU are assumed to also apply to the unsampled DU and decisions are made accordingly. ****This approach may not be acceptable to individual states or other regulatory agencies based on feedback received during the comment period and the possible introduction of decision errors (See Section 4.4.4.2)****

Narrative

� Frequently proposed for large tracts of land or large volumes of soil to reduce costs.

•May be necessary when funds or other constraints prohibit sampling all DUs

� ISM has no unique capability to enable extrapolation or reduce the uncertainty involved.

◦ The issue is the same whether the DU is sampled by ISM or with discrete sampling – is there sufficient basis to assume that the mean concentration estimated there is the same (or higher) than the mean concentration in unsampled DUs.

� Must assume that both the mean and the distribution of contaminants are essentially equivalent across DUs

87

88


Making Decisions

Decision Mechanisms

Data Evaluation

Speaker bullets – decision mechanisms decided upfront / data evaluation

The purpose of this section is to describe how to evaluate the data to ensure that sampling objectives have been met.

88

89

Data Evaluation Components

Data EvaluationInterpreting error

Identifying sources of error

Quantifying error

Isolating sources of error

Data evaluation can include several steps that we will discuss in the following slides

89

90

Identifying Sources of Error

Field Number of increments

Increment collection

Field processing

Field splitting

DU size and shape

Laboratory Lab processing

Subsampling

Extraction

Digestion

Analysis

Error (Variability) – not “mistake”

Speaker Bullets – not comprehensive; theoretical sources of error discussed in Chapter 2 (Section 2.2) ISM controls these better than discrete

It is useful when evaluating data to consider the steps in the process where errors (variability) may have been introduced

If error is determined to be unacceptable in a given data set – one or more of these sources may be at fault

90

91

Quantifying Error

Decision Unit

Data includes all sources of error

RSD = CV = standard deviation / arithmetic mean

ITRC, ISM-1, Section 4.3.1.3 and Section 7.3

Narrative - Evaluation of overall error can by accomplished by using replicate data to calculate a relative standard deviation (RSD) for the replicates. The RSD is a statistical measure of variance that quantifies the precision of the data, although not the accuracy of the estimate. RSD is a measure of reproducibility of estimates of the mean provided by replicates. It provides a measure of the total error associated with the replicates. (see Section 4.3.1.3 for more info)

In order to calculate appropriate statistics, at least three field replicate samples are needed.

91

92

Interpreting Error

“Unacceptable” RSD

Low RSD

High RSD

ITRC, ISM-1, Section 4. 3.4.4 and Section 7.3

Unbiased Biased

Imprecise

Precise

• While there are rules of thumb provided in other guidance documents (CHECK WITH YOUR STATE REGULATORY AGENCY), the ITRC offers no prescribed definition of “acceptable” RSD (should be evaluated on a site-by-site basis) – See Section 7.3 for more info

•Low RSD – Generally you don’t need to go back and identify sources of error. While A low RSD may indicate that the field replicates are providing reproducible estimates of the average, this is not always the case -- is not always an indication that the mean is accurate or that the 95% UCL exceeds the population mean (unless the distribution can be reasonably assumed to be relatively homogeneous) ** An estimate of the average from replicates with a low RSD is not necessarily close to the actual mean. The opportunity for significant error is greatest when the DU is relatively heterogeneous and the replicates by chance give similar results.

•High RSD may indicate errors introduced through lab or field activities or may be attributable to heterogeneity in site concentrations. Generally leads to the need to isolate potential sources of error –discussed on the earlier slide “Identifying Sources of Error” (see Section 4.3.4.4 for more info)

•Lab

•Field

RSD measures reproducibility, not accuracy of the results (see Section 4 for addition discussion of these concepts).

If the goal is to make sure that the mean is not underestimated, a 95UCL should be calculated regardless of whether the RSD is high or low.

92

93

Isolating Sources of Error

Adapted from EPA 2011, page 38: http://go.usa.gov/EAE

Ideally, the project team will then designate one of these replicates for separation into laboratory replicates.) Replicate RSD data is intended to quantify the total error of the measurement system; and if one sample is split into laboratory replicates then the total error can be attributed to either field sampling or laboratory procedures.

93

94


Making Decisions

Decision Mechanisms

Data Evaluation

Speaker bullets –The purpose of this section is to describe how to use the data generated through ISM samples to make decisions regarding a DU and how to evaluate the data to ensure that sampling objectives have been met.

Narrative: ISM provides estimates of the mean concentration of the contaminant in one or more SUs that can be used to make a decision regarding the DU. Since the data may inform one or more decision objectives, it is helpful to establish a structured approach to making decisions, referred to here as decision mechanisms. Systematic evaluation of the data can determine if sampling errors have been adequately addressed for the decisions being made.

94

95


Plan

Principles

Systematic Planning

Statistical Design

Implement


Lab Processing



Application

Part 1

Part 2

Section 2

Section 3

Section 4

Section 5

Section 7

Section 8

Section 6


ISMOpportunities

In the next few slides we will look at some of the types of sites and applications where ISM may be a useful tool to consider, briefly touch on cost consideratinos and then we’ll recap some of the take home messages about ISM that we’d like for you to leave with today.

95

96

ISM Applications

Regulated sites

Residential yards

Former pesticide-applied orchards

Stockpiled soil

Post-soil treatment sampling

Dredged materials

Residential Yard

Soil Stockpile

Other Applications

We recognize that there are many potential applications for ISM in the work we do. We have focused on regulated sites – and there are already examples of where ISM has been applied successfully, with much thanks to members of the ITRC’s ISM Technical Team.

However, we recognize that there can be reluctance and resistance to implementing ISM on regulated sites, so perhaps you can think of other opportunities – other circumstances and types of sampling – where there may be an easier ‘point of entry’ for demonstrating ISM in your state/region/jurisdiction. These (applications) are just some of the opportunities identified by members of the ISM Team - instances where the mean is appropriate & ISM makes sense (more so than discrete sampling).

Applications are grouped together in the following manner – evenly dispersed discharges, mixed/lessvariable soils, large areas, etc., and continue on to the next slide.

In NJ and many other states, there are restrictions (regulations) which limit the use of composite soil sampling as well as the averaging of soil sample results. Please refer to Table 8-1 in Section 8 of the ISM document for a complete listing of all states with restrictions on compositing. In NJ, however, there is a process whereby one can petition for a "variance" from the technical requirements and guidance.

The rationale for varying from the requirement needs to include documentation that the work conducted has achieved the objectives of the rule requirement or guidance from which it varied. Thus ISM could be utilized in NJ, and possibly other states, by providing the technical justification that it not only meets but most likely exceeds the discrete soil sampling requirements as well as its objectives.

Speaker Note: We realize there are concerns…….others have had them too…….but open yourself to an opportunity for better decision- making through ISM

96

97

ISM Applications (continued)

Large Areas

• Tailings impoundments

• Agricultural fields

• Floodplain soils

Transects

Dredged materials

Mine Tailings Impoundment

Transects

2 miles

Discrete VS. ISM

DU1

DU2

DU3

DU4

DU5

Dis

tanc

e

Mine Tailings

Other Applications

We recognize that there are many potential applications for ISM in the work we do. We have focused on regulated sites – and there are already examples of where ISM has been applied successfully, with much thanks to members of the ITRC’s ISM Technical Team.

However, we recognize that there can be reluctance and resistance to implementing ISM on regulated sites, so perhaps you can think of other opportunities – other circumstances and types of sampling – where there may be an easier ‘point of entry’ for demonstrating ISM in your state/region/jurisdiction. These (applications) are just some of the opportunities identified by members of the ISM Team - instances where the mean is appropriate & ISM makes sense (more so than discrete sampling).

Applications are grouped together in the following manner – evenly dispersed discharges, mixed/lessvariable soils, large areas, etc., and continue on to the next slide.

In NJ and many other states, there are restrictions (regulations) which limit the use of composite soil sampling as well as the averaging of soil sample results. Please refer to Table 8-1 in Section 8 of the ISM document for a complete listing of all states with restrictions on compositing. In NJ, however, there is a process whereby one can petition for a "variance" from the technical requirements and guidance.

The rationale for varying from the requirement needs to include documentation that the work conducted has achieved the objectives of the rule requirement or guidance from which it varied. Thus ISM could be utilized in NJ, and possibly other states, by providing the technical justification that it not only meets but most likely exceeds the discrete soil sampling requirements as well as its objectives.

Speaker Note: We realize there are concerns…….others have had them too…….but open yourself to an opportunity for better decision- making through ISM

97

98

ISM Applications (continued)

Firing Ranges

Confirmatory sampling

Background

Other

• Fill material

• “Rail to trail” sites

Small Arms Firing Range

Post-Excavation Confirmatory Sampling

SLIDE on ISM Applications (continued)

So maybe you have already seen some opportunities – either on regulated sites or some other circumstance – where ISM might be a good fit. Or maybe you could bring up some other possibilities not listed here during the question/answer period that’s coming up.

Another tool in the toolbox. Remember ISM, although an improvement in regards to environmental sampling doesn’t invalidate what we have been doing…

….but we think it will make life a whole lot easier (and better) in many, many circumstances.

98

99

How Does ISM Cost Compare?

Elements

Planning

Field Collection

QA/QC Samples

Sample Transport

Sample Processing/Conditioning

Lab Analysis

Overall Sampling/Analysis Portion of Project

• It depends

• We can pull out each element of a sampling/analysis project and take a look at how ISM might compare

• Planning – about the same, maybe a little more for those less familiar with systematic planning and DU concept

• Field Collection – Often ISM costs more

• QA/QC – can be fewer QA/QC samples with ISM since fewer overall samples collected

• Sample Transport can be higher with ISM since samples are so much larger

• Sample Processing/Conditioning – since there often is none of this with discrete, ISM will add to cost here.

• Lab analysis – per sample can be higher due to subsampling protocol; but fewer samples submitted can mean lower lab bill

• Overall – often the effect of fewer samples collected for analysis means a lower overall cost

• But remember, these are only sampling/analysis costs; as we’ll see in a couple slides that does not really tell the whole story

• But first, let’s look at a specific cost comparison example

99

100

A Cost Comparison Example

US Army Corps of Engineers Study Metals in Soil

Activity

Per Sample Cost($)

Total Project Sampling/Analysis Cost ($)

ISM Discrete ISM1 Discrete2 Discrete3

Field Sampling 35-50 10-15 105-150 70-105 150-225

Lab Prep 40-60 0-10 120-180 0-70 0-150

Analysis 225-275 125-135 675-825 875-945 1,875-2,025

Total 300-385 135-160 945-1,155 945-1,120 2,025-2,400

1 Based on 3 replicate 100-increment ISM/DU2 Based on collection of 7 discrete samples/DU3 Based on collection of 15 grab samples/DU

Source: US Army Corps, Cost and Performance Report of Incremental Sampling Methodology for Soil Containing Metallic Residues, ERDC TR-13-10, September 2013

•Two case studies in guidance doc looked at cost - both showed 50%-80% savings vs. discreet; both looked at energetics

•This table is from a recent USACE cost comparison study looking at metals in soil

•Included two discrete sampling scenarios; 7 samples/DU and 15/DU

•Found that although per sample cost 35-53% higher, overall project cost was up to 40% lower.

•Again want to emphasize these are just the sampling analysis project costs

100

101

Bottom Line on Cost Comparisons

Measuring the cost difference between ISM and discrete sampling.

Measuring the cost of making a wrong decision.

•Sampling/analysis usually leads to some decision

•Quality of the decision is only as good as quality of the data on which it is made

•We’ve seen how ISM can provide a very high quality estimate of mean concentration

•So while it is relatively easy to compare cost of ISM vs. discrete sampling/analysis plans,

•Ultimately what matters is the cost of making a decision error – how expensive is it to unnecessarily excavate 100 yd3 soil? How expensive is it to walk away from a site that actually poses a risk to human health?

101

102

ITRC’s ISM Solution

Plan

Principles

Systematic PlanningStatistical

Design

ImplementField

ImplementationLab

Processing


Web-Based Document at: http://www.itrcweb.org/ISM-1/

Here’s our answer to how to get a representative sample.

This flow chart/illustration provides a snapshot of the contents and structure of our two-day training – and of our Technical and Regulatory Guidance Document on incremental sampling methodology. From planning through assessment we’ve developed a process for considering all the important aspects of getting that ‘representative sample’ we’ve been talking about. And remember – all the elements we have displayed here in this figure would pertain to other types of sampling – but it turns out that incremental sampling does a better job for many circumstances, and that is the focus of our document and training.

The Technical and Regulatory Guidance Document is over 400 pages (web-based) – so hopefully this 2-session training has given you the comfort - and motivation - to dive in to the more detailed presentation in the actual document.

102

103

Overview & Wrap-up

Specifically designed to address short comings of discrete sampling methods

Provides a systematic, science-based approach to site investigation

Increases data representativeness

Provides greater confidence in decision making

ISM:


103

104

104

Thank You

2nd question and answer break

Links to additional resources• http://www.clu-in.org/conf/itrc/ISM/resource.cfm

Feedback form – please complete• http://www.clu-in.org/conf/itrc/ISM/feedback.cfm

Follow ITRC

Need confirmation of your participation today?

Fill out the feedback form and check box for confirmation email and certificate.

Links to additional resources:

http://www.clu-in.org/conf/itrc/ISM/resource.cfm

Your feedback is important – please fill out the form at:

http://www.cluin.org/conf/itrc/ISM/feedback.cfm

The benefits that ITRC offers to state regulators and technology developers, vendors, and consultants include:

Helping regulators build their knowledge base and raise their confidence about new environmental technologies

Helping regulators save time and money when evaluating environmental technologies

Guiding technology developers in the collection of performance data to satisfy the requirements of multiple states

Helping technology vendors avoid the time and expense of conducting duplicative and costly demonstrations

Providing a reliable network among members of the environmental community to focus on innovative environmental technologies

How you can get involved with ITRC:

Join an ITRC Team – with just 10% of your time you can have a positive impact on the regulatory process and acceptance of innovative technologies and approaches

Sponsor ITRC’s technical team and other activities

Use ITRC products and attend training courses

Submit proposals for new technical teams and projects